The purpose of this study was to determine the effect of sampling strategy (i.e., number of breaths) on measured peak rate of oxygen uptake ( VO(2peak)) elicited by a range of severe intensity exercise bouts. The hypothesis was that a smaller sample (i.e., fewer breaths) would produce a higher measure of VO(2peak) and that this effect would be greater in shorter tests than in longer tests. Thirty-three university students performed constant-power cycle ergometer tests at intensities selected to elicit fatigue in ~3.0 min (short duration), approximately 5.5 min (medium duration), and approximately 8.0 min (long duration). Values for VO(2peak) were the highest rates of oxygen uptake obtained using the following sampling methods: single breath, and 3-, 5-, 15- and 30-breath rolling averages. As hypothesized, measures of VO(2peak) increased systematically with decreasing sample size. Contrary to the hypothesis, the effect of sample size was greater in medium duration and long duration tests than in the short duration tests. The interaction between test duration and sample size on measures of VO(2peak) highlights the importance of standardizing the analysis protocol for exercise in the severe domain. If such standardization is not feasible, it should be recognized that specific analysis protocols may exert a substantial effect upon the reported VO(2peak).